Back row passenger safety monitoring system and method based on face recognition

Abstract

The invention discloses a back-row passenger safety monitoring system and method based on face recognition, the system comprises a camera, a cabin area controller and a reminding module, the camera is used for collecting passenger face image data on back-row seats, and the output end of the camera is connected to the cabin area controller; and the cabin area controller recognizes the state of the passengers on the back row based on the face image data and sends reminding information to the driver through the reminding module based on the state of the passengers on the back row. Real-time monitoring of rear-row passengers in the driving process of an automobile is achieved through the camera, the large screen, the face recognition system and the like, when the rear-row passengers are abnormal, the abnormal conditions are fed back to a driver in time, and therefore the safety of the rear-row passengers is improved.

Description

Technical Field

[0001] The present invention relates to the field of automobile safety monitoring interaction, and in particular to a rear passenger safety monitoring system and method based on facial recognition. Background Art

[0002] With the development of the automobile industry, the application of intelligent technology in vehicles has realized the reliability and safety of vehicles. However, the existing technology often only focuses on the status of the driver in front, such as driver fatigue, etc. For example, a driver fatigue status recognition method with patent application number 201810733592.6 includes: collecting at least one of the driver image information, driver operation information and vehicle driving information; based on the collected information, preliminarily identifying the driver's fatigue status and obtaining the preliminary recognition result of the driver's fatigue status; changing the vehicle driving status according to the preliminary recognition result of the driver's fatigue status and the driving scene of the vehicle; further determining the driver's fatigue status result according to the driver's operation data for adjusting the vehicle driving status and the corresponding vehicle driving data. Accurate recognition of the driver's fatigue status is achieved.

[0003] As shown in the above-mentioned public documents, the existing technology often only focuses on the driver's status, whether the driver is tired or not, and often ignores the status of the rear passengers and the interaction between the front driver and the rear passengers. During the driving of the car, the rear passengers may suddenly fall ill, but the front driver cannot pay attention to the condition of the rear passengers in time. The front driver is limited by the blind spot of vision and the need for driving safety and cannot pay attention to the status of the rear passengers in time. When the rear passengers suddenly become aphasic or faint due to illness, many irreversible injuries, serious and even life-threatening, occur because they are not discovered and treated in time. The existing technology does not have a good monitoring method for the status of the rear passengers. The driver cannot obtain the status of the rear passengers in time, and cannot handle and interact in time to improve the safety monitoring of the rear passengers of the vehicle. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a rear passenger safety monitoring system and method based on facial recognition. The user monitors the safety status of the rear passengers by performing facial recognition through the rear camera acquisition and promptly reminds the front driver to pay attention, thereby improving the safety of the rear passengers.

[0005] In order to achieve the above-mentioned purpose, the technical solution adopted by the present invention is: a rear passenger safety monitoring system based on facial recognition, including a camera, a cabin domain controller, and a reminder module. The camera is used to collect facial image data of passengers on the rear seats, and its output end is connected to the cabin domain controller; the cabin domain controller identifies the status of the rear passengers based on the facial image data and sends a reminder message to the driver through the reminder module based on the status of the rear passengers.

[0006] The cameras are multiple cameras arranged on the driver and co-driver seats and facing the rear row.

[0007] The cockpit domain controller obtains facial feature point information of rear passengers through face recognition based on facial image data, and determines the abnormal state of the rear passengers based on the facial feature point information of the rear passengers and outputs a control signal matching the abnormal state.

[0008] The reminder module includes a large vehicle screen, and the cockpit domain controller drives and controls the large vehicle screen to display the rear row image taken by the camera and remind the driver through voice that the rear row passengers are in an abnormal state.

[0009] A face recognition algorithm is running in the cockpit domain controller, which sequentially obtains facial feature point information of the rear-seat passengers through face monitoring, face alignment, image preprocessing, and feature extraction steps, and then compares the facial feature point information of the rear-seat passengers with pre-stored facial feature point information of normal expressions to determine whether the rear-seat passengers are abnormal at this time. When the abnormality lasts for a set time, it is determined that the rear-seat passengers are abnormal and a corresponding control signal is issued.

[0010] The system also includes a call button, which is connected to the cockpit domain controller, and the cockpit domain controller is connected to the vehicle phone module. After the call button is triggered, the vehicle phone module is driven to call 120.

[0011] The monitoring system also includes a rear passenger position detection module, which is used to detect whether the rear passengers are within a set safety range. Its output end is connected to the cockpit domain controller. After the cockpit domain controller detects that the rear passengers are beyond the safety range and the duration reaches a set time threshold, the cockpit domain controller displays the delayed alarm time through the on-board large screen and dials the alarm number after the delay ends. At this time, a cancel alarm button is provided through the on-board touch screen or hardware button. If no cancellation signal is received after the delay is completed, the alarm number is directly dialed or an alarm text message is sent.

[0012] The cockpit domain controller is provided with a dedicated memory for storing face recognition data of rear passengers. The face recognition data is encrypted and saved and deleted after a set time.

[0013] A rear passenger safety monitoring method based on facial recognition, comprising:

[0014] The facial image data of the passengers in the rear seats is collected through the camera; the cockpit domain controller identifies the status of the rear passengers based on the facial image data and sends a reminder message to the driver through the reminder module based on the status of the rear passengers; when it is identified that the rear passengers are in an abnormal state, the cockpit domain controller drives the large-screen display of the vehicle to display the rear images taken by the camera and reminds the driver through voice that the rear passengers are in an abnormal state.

[0015] After the driver learns of the abnormal state of the rear passenger through the in-vehicle large screen and voice, the cockpit domain controller monitors in real time whether the call button is triggered. When the call button is triggered, it drives the in-vehicle phone module to call 120.

[0016] The advantages of the present invention are: real-time monitoring of rear passengers during driving of the car is achieved through cameras, large screens, facial recognition systems, etc., and timely feedback is given to the driver when abnormalities occur in the rear passengers, thereby improving the safety of the rear passengers; at the same time, in order to avoid risks such as driver's safety authorization robbery, the activity range of the rear passengers is defined, and when the activity range of the rear passengers exceeds the set range and lasts longer than the set time, a reminder is issued to the front driver, and an alarm is automatically issued to avoid robbery and the like, thereby improving the safety of the front driver; at the same time, the abnormal recognition data of the rear passengers is only used to remind the driver, so their facial data is encrypted and stored and deleted after a delay to avoid privacy leakage. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The following is a brief description of the contents expressed in the drawings of the present invention and the symbols in the drawings:

[0018] Figure 1 A simplified flow chart of a rear passenger safety monitoring method based on facial recognition;

[0019] Figure 2 A control logic diagram of a rear passenger safety monitoring method based on facial recognition;

[0020] Figure 3 Schematic diagram of facial feature points;

[0021] Figure 4 Schematic diagram of eye feature points. DETAILED DESCRIPTION

[0022] The specific implementation of the present invention will be further explained in detail below by describing the optimal embodiment with reference to the accompanying drawings.

[0023] The main technical problem solved by this application is that the rear passengers are not easy to be observed by the front driver. When the rear passengers are abnormal, the front driver cannot get timely information, which causes potential safety hazards. The rear passenger abnormalities here include but are not limited to abnormalities caused by physical reasons of the rear passengers, abnormalities caused by active behaviors such as rear passengers jumping out of the window and robbing. The specific solution is as follows:

[0024] In this embodiment, a rear passenger safety monitoring system based on facial recognition includes a camera, a cockpit domain controller, and a reminder module. The camera is used to collect facial image data of passengers in the rear seats, and its output end is connected to the cockpit domain controller; the cockpit domain controller identifies the status of the rear passengers based on the facial image data and sends a reminder message to the driver through the reminder module based on the status of the rear passengers.

[0025] In order to fully cover the images of the rear passengers, clear facial images are acquired through multi-directional image acquisition, so the cameras are multiple cameras facing the rear row, which are arranged on the driver and co-driver seats. Preferably, two cameras are arranged, which are respectively arranged on the headrests of the driver and co-driver seats and facing the rear passengers, so that the images of the rear passengers can be acquired, and then the facial image data can be acquired. In a preferred solution, in order to better capture the facial data, the camera can be set to be movable, so that the shooting angle of the camera can be adjusted by Huadong to realize the acquisition of facial images, and the acquired facial image data is sent to the cockpit domain controller for execution.

[0026] The cockpit domain controller obtains the facial feature point information of the rear passengers through face recognition based on the facial image data, and judges the abnormal state of the rear passengers based on the facial feature point information of the rear passengers and outputs a control signal matching the abnormal state. The cockpit domain controller is an on-board controller, which is a computing platform specially used for the control and management of electronic equipment inside the car. It has high-performance computing capabilities and can support high-load tasks such as high-definition video and 3D graphics. The cockpit domain controller usually has high computing power and graphics processing capabilities. Therefore, this solution uses the cockpit domain controller to run the face recognition algorithm and the corresponding control logic to achieve control. The cockpit domain controller analyzes and identifies the collected face data to determine whether the facial state of the rear passengers is abnormal at this time. The abnormality includes facial painful expressions, etc. If the recognition is judged to be abnormal facial expression, a corresponding control signal is issued. The control signal is used to drive the reminder module to issue a corresponding reminder signal of the abnormality of the rear passengers. The reminder module is used to remind the driver of the abnormal state of the rear passengers.

[0027] The reminder module includes a large in-vehicle screen and an in-vehicle voice system. The cockpit domain controller drives and controls the large in-vehicle screen to display the rear-seat images taken by the camera and reminds the driver through voice that the rear-seat passengers are in an abnormal state. After recognizing that the rear-seat passengers are abnormal, since the driver can only see the rear-seat passengers through the rearview mirror and the viewing angle is limited, after the system recognizes that the rear-seat passengers are abnormal, the cockpit domain controller controls the large in-vehicle screen to switch the display screen to the collected abnormal facial expressions, and displays the faces of the rear-seat passengers to the user through the large in-vehicle screen. At the same time, the driver is reminded of the abnormality of the rear-seat passengers through voice reminders. In this way, the voice can remind the driver to pay attention, and the large in-vehicle screen in the front row can facilitate the driver to view the faces of the rear-seat passengers. In this way, it is convenient for the front-row driver to pay attention to the status of the rear-seat passengers in a timely manner, avoiding safety hazards and accidents caused by abnormal conditions such as sudden illness of the rear-seat passengers that cannot be detected by the driver in time.

[0028] In order to realize the recognition of facial abnormalities, this solution runs a face recognition algorithm in the cockpit domain controller. The face recognition algorithm obtains the facial feature point information of the rear passengers through the steps of face monitoring, face alignment, image preprocessing, and feature extraction. Then, the facial feature point information of the rear passengers is compared with the pre-stored normal facial feature point information to determine whether the rear row is abnormal at this time. When the abnormality lasts for a set time, the rear passengers are judged to be abnormal and a corresponding control signal is issued. The facial feature point information of the normal face is stored in the database in advance, and then the facial expression feature point information of the identified rear passengers is compared to determine whether it is normal; if it is normal, no operation is performed, otherwise, the face is judged to be abnormal, and the cockpit domain controller issues an abnormal reminder through the large screen and voice.

[0029] A soft switch for dialing 120 with one button is developed in the large screen of the vehicle, or a button for dialing 120 is set, the button for dialing is connected to the cockpit domain controller, the cockpit domain controller is connected to the vehicle phone module, and after the button for dialing is triggered, the vehicle phone module is driven to dial 120. When the driver in the front row is reminded that the rear passenger has an abnormality, the driver can dial 120 emergency services with one button by triggering the switch, thereby achieving the purpose of rapid rescue.

[0030] In this solution, since the facial information of the rear passengers will be obtained in real time, in order to avoid security risks such as privacy leakage, the cockpit domain controller is equipped with a dedicated memory, which is used to store the facial recognition data of the rear passengers. The facial recognition data is encrypted and saved and deleted after a set time. That is, when the facial image data collected by the camera is sent to the cockpit domain controller for processing, the image is stored in the dedicated memory at the same time, and the storage method is encrypted storage; after the storage reaches the set time, the image in the dedicated memory is deleted. The purpose of doing this is to facilitate the call of the image after a delay to display on the front car large screen. The main purpose of encrypted storage and deletion is to avoid the leakage of private facial information and protect the information of passengers.

[0031] In a preferred solution of this embodiment, the above rear passenger monitoring is only for monitoring when the rear passenger is ill, etc. However, in the prior art, there may be situations where the rear passenger interferes with the driver's driving for various reasons, such as robbery, disputes, etc. Therefore, in order to achieve the monitoring of the rear passenger interfering with the front driver, a rear passenger position detection module is provided, which is used to detect whether the rear passenger is within the set safety range. Its output end is connected to the cockpit domain controller. After the cockpit domain controller detects that the rear passenger exceeds the safety range and the duration reaches the set time threshold, the cockpit domain controller displays the delayed alarm time through the vehicle-mounted large screen and dials the alarm phone after the delay ends. At this time, a cancel alarm button is provided through the vehicle-mounted touch screen or hardware button. If the cancellation signal is not received after the delay is completed, the alarm phone is directly dialed or an alarm text message is sent. The rear passenger position detection module uses image recognition to identify the position range of the passenger, and judges whether the rear passenger exceeds the normal safety range based on image recognition. When the rear passengers are riding, their position is fixed. The safe activity range of the rear passengers is marked in advance through calibration. When performing image recognition, the system collects, analyzes and identifies in real time whether the current rear passengers are beyond the set safe position range. If so, timing is started. When the rear passengers are beyond the safe range and the duration reaches the set time threshold, the cockpit domain controller determines that the rear passengers pose a safety threat to the front driver's driving. Therefore, at this time, the cockpit domain controller drives the on-board large screen to display the delayed alarm time and dial the alarm number after the delay ends. The on-board large screen displays that an alarm is about to be issued and a countdown delay is performed. During the delay period, if the driver operates the cancel button, the alarm is canceled. Otherwise, after the timing ends, the cockpit domain controller drives the on-board phone module to directly dial the 110 alarm number or send an alarm message to realize direct alarm, thereby reducing or warning the impact of the rear passengers on the safe driving of the driver in front, and at the same time, the problem can be solved in time through the police. Regarding the setting of the safe position range, the image data collected by the camera can be divided into the front area and the rear area in a pre-calibrated manner. If the camera recognizes that a passenger enters the front area from the rear area, it is judged that the passenger has exceeded the safe area. The rear area here is the set safe position range for the rear passengers, which is equivalent to setting an electronic fence range for the rear passengers. Image recognition is used to identify whether the rear passengers have exceeded the electronic fence range. If not, it is judged to be normal; if they exceed the electronic fence range and the duration is greater than the set time threshold, it is judged that the rear passengers pose a safety threat to the front driver's driving, and the 110 alarm dialing process is initiated, that is, the cockpit domain controller drives the on-board large screen to display the delayed alarm time and dials the alarm number after the delay ends.

[0032] When it is identified that the rear passengers pose a safety threat to the front driver's driving, the cockpit domain controller is connected to the power domain controller, which controls the vehicle speed to gradually reduce linearly until it stops. At the same time, when reducing the speed, the cockpit domain controller controls the vehicle's hazard warning lights (double flash lights) to turn on to remind passing vehicles to pay attention to safety. If the vehicle has ADAS assisted driving function, the ADAS assisted driving function will be called to control the vehicle to change lanes and stop after pulling over while the vehicle is reducing its speed, thereby improving safety.

[0033] In another preferred embodiment of the present application, in order to avoid safety hazards caused by quarrels between rear passengers and front passengers, such as jumping out of the window and other safety accidents, the quarrel sound between the front and rear passengers is identified by a voice recognition system, and the tone and semantic analysis of the front and rear passengers' speech are used to determine whether the front and rear passengers are quarreling, or the image face recognition is used to analyze whether the front driver and / or the rear passengers are angry. When it is determined that either the front or rear passenger or the driver is angry, the vehicle speed is detected in real time, and when the vehicle speed is greater than a set speed threshold, the door is locked; then the window status is monitored in real time, and if the window is in an open state, the window is closed to a small opening state, such as the window opening degree is less than half of the full opening; otherwise, the door handle and window switch button are monitored in real time to determine whether they are triggered. If so, it means that there are safety hazards such as rear passengers opening the door and window too much during driving of the vehicle. At this time, after the trigger signal is detected, all the windows are directly closed to improve the safety of the vehicle. At this time, the window and door control functions are prohibited and operation is not allowed. When the vehicle speed is lower than the set threshold, all windows are forced to close, the window opening or closing function is activated and allowed to be operated by rear passengers. This can effectively improve the driving safety of the vehicle, avoid irrational behaviors such as jumping out of the window or opening the car door due to quarrels, improve safety, and automatically release the restrictions when the vehicle speed is lower than the set value, improving safety.

[0034] This embodiment also provides a rear passenger safety monitoring method based on facial recognition, including: collecting facial image data of passengers on the rear seats through a camera; the cockpit domain controller identifies the status of the rear passengers based on the facial image data and sends a reminder message to the driver through a reminder module based on the status of the rear passengers; when it is identified that the rear passengers are in an abnormal state, the cockpit domain controller drives and controls the on-board large screen to display the rear image taken by the camera and reminds the driver through voice that the rear passengers are in an abnormal state.

[0035] After the driver learns of the abnormal state of the rear passenger through the in-vehicle large screen and voice acquisition, the cockpit domain controller monitors in real time whether the dial phone button is triggered, and drives the in-vehicle phone module to dial 120 after the dial phone button is triggered. A one-touch dial 120 soft switch is developed in the in-vehicle large screen or a one-touch dial 120 button is set. The dial phone button is connected to the cockpit domain controller, and the cockpit domain controller is connected to the in-vehicle phone module. After the dial phone button is triggered, the driver can dial 120 emergency by triggering the switch, thereby achieving the purpose of rapid rescue.

[0036] The facial feature point information of the rear passengers is compared with the pre-stored normal facial feature point information to determine whether the rear seats are abnormal at this time. When the abnormality lasts for a set time, the rear passengers are judged to be abnormal and a corresponding control signal is issued. The facial feature point information of normal faces is stored in the database in advance, and then the facial feature point information of the identified rear passengers is compared to determine whether it is normal; if it is normal, no operation is performed, otherwise the face is judged to be abnormal, and the cockpit domain controller issues an abnormal reminder through the large screen and voice.

[0037] In this solution, since the facial information of the rear passengers will be obtained in real time, in order to avoid security risks such as privacy leakage, the cockpit domain controller is equipped with a dedicated memory, which is used to store the facial recognition data of the rear passengers. The facial recognition data is encrypted and saved and deleted after a set time. That is, when the facial image data collected by the camera is sent to the cockpit domain controller for processing, the image is stored in the dedicated memory at the same time, and the storage method is encrypted storage; after the storage reaches the set time, the image in the dedicated memory is deleted. The purpose of doing this is to facilitate the call of the image after a delay to display on the front car large screen. The main purpose of encrypted storage and deletion is to avoid the leakage of private facial information and protect the information of passengers.

[0038] This solution uses a camera to collect the situation in the back row for facial recognition, actively monitors the status of the back row passengers, and gives the driver timely feedback, which can greatly improve the safety of the back row passengers and the speed of disease discovery and treatment. Through cameras, large screens, facial recognition systems, etc., the car can monitor the back row passengers in real time while driving, and provide timely feedback to the driver when the back row passengers are abnormal, thereby improving the safety of the back row passengers. The solution specifically includes the following steps:

[0039] (1) A camera is placed at the head of the driver and co-driver seat to obtain information about the rear passengers, and the captured images are transmitted to the cockpit domain controller for processing. The facial feature point information of the rear passengers is obtained through a four-step algorithm of face monitoring, face alignment, image preprocessing, and feature extraction. When the passenger is in a state of fainting or pain, the facial expression changes, and the eye feature points exceed the normal threshold. The domain controller sends a danger signal to the rear passengers, the large screen switches the camera angle to reflect the rear image, and sends a voice to remind the driver that the rear passengers are in a dangerous state.

[0040] (2) After the driver confirms that everything is correct, 120 is automatically dialed with one click.

[0041] (3) The memory has pre-stored facial feature point information of normal expressions, and the controller will send a signal when it detects that the feature point information of the rear passenger is abnormal for a period of time.

[0042] The present invention mainly involves a cockpit domain controller, a large screen, a camera, etc. The specific control method is as follows Figure 2 shown.

[0043] (1) Opening and closing the function:

[0044] The rear safety monitoring mode can be added through the large screen and I HU (display control unit) to turn on / off the function. After the function is turned on, the large screen transmits the signal to the cockpit domain controller, which then wakes up the camera configured at the head of the driver and co-driver seat, and the entire rear passenger safety monitoring function starts working to monitor the status of the rear passengers in real time.

[0045] (2) Specific implementation of functions:

[0046] The camera takes pictures to obtain pictures with facial information, and the pictures are transmitted to the domain controller for processing. Figure 1 deal with.

[0047] Face detection and positioning: Face detection is mainly used for processing before face recognition in practice, that is, to determine whether there is a face in the image, and if so, to accurately calibrate the position and size of the face. Find the position of the face in the image, take the upper left corner of the image as the coordinate origin, record the coordinates of the upper left corner and lower right corner of the face frame, and crop the face part. The subsequent algorithm will process the face part.

[0048] Face image preprocessing: Since the original image obtained by the camera is often not directly usable due to various restrictions and random interference, it must be preprocessed with grayscale correction, noise filtering, etc. in the early stages of image processing, and ultimately serve the feature point extraction process. This step will obtain an image that can be used to extract feature points.

[0049] Feature point extraction: There are 68 feature points in total on the face (such as Figure 3 ), we can apply facial landmark detection to locate the facial area, including eyes, eyebrows, nose, ears and mouth, which also means that we can extract the required facial structure through the required facial area, including the human eye area we need.

[0050] The eye aspect ratio (EAR) is calculated from the eye landmarks. Figure 4 The 6 feature points p1 to p6 in are the 6 feature points corresponding to the eyes in the facial feature points. Each eye is represented by 6 (x, y) coordinates. Based on this description, the equation of EAR can be derived:

[0051]

[0052] Get the aspect ratios of the two eyes EAR1 and EAR2 and compare them with the threshold

[0053] Compare EAR1 and EAR2 with the normal thresholds L and H stored in the memory.

[0054]

[0055] When EAR1 or EAR2 exceeds the threshold, the domain controller sends an abnormal signal to the IHU control screen to display the rear image and issue a language reminder. Two options of abnormality / misjudgment pop up for the driver to choose. When the driver confirms the abnormality, call 120. When the driver confirms the misjudgment, turn off the function.

[0056] Obviously, the specific implementation of the present invention is not limited to the above-mentioned methods. As long as various non-substantial improvements are made using the method concept and technical solution of the present invention, they are all within the protection scope of the present invention.

Claims

1. A rear passenger safety monitoring system based on facial recognition, characterized in that: It includes a camera, a cockpit domain controller, and a reminder module. The camera is used to collect facial image data of passengers in the rear seats, and its output end is connected to the cockpit domain controller; the cockpit domain controller identifies the status of the rear passengers based on the facial image data and sends a reminder message to the driver through the reminder module based on the status of the rear passengers.

2. A rear passenger safety monitoring system based on facial recognition as claimed in claim 1, characterized in that: The cameras are multiple cameras arranged on the driver and co-driver seats and facing the rear row.

3. A rear passenger safety monitoring system based on facial recognition as claimed in claim 1 or 2, characterized in that: The cockpit domain controller obtains facial feature point information of rear passengers through face recognition based on facial image data, and determines the abnormal state of the rear passengers based on the facial feature point information of the rear passengers and outputs a control signal matching the abnormal state.

4. A rear passenger safety monitoring system based on facial recognition as claimed in claim 1 or 2, characterized in that: The reminder module includes a large vehicle screen, and the cockpit domain controller drives and controls the large vehicle screen to display the rear row image taken by the camera and remind the driver through voice that the rear row passengers are in an abnormal state.

5. A rear passenger safety monitoring system based on facial recognition as claimed in any one of claims 1 to 4, characterized in that: A face recognition algorithm is running in the cockpit domain controller, which sequentially obtains facial feature point information of the rear-seat passengers through face monitoring, face alignment, image preprocessing, and feature extraction steps, and then compares the facial feature point information of the rear-seat passengers with pre-stored facial feature point information of normal expressions to determine whether the rear-seat passengers are abnormal at this time. When the abnormality lasts for a set time, it is determined that the rear-seat passengers are abnormal and a corresponding control signal is issued.

6. A rear passenger safety monitoring system based on facial recognition as claimed in any one of claims 1 to 4, characterized in that: The system also includes a call button, which is connected to the cockpit domain controller, and the cockpit domain controller is connected to the vehicle phone module. After the call button is triggered, the vehicle phone module is driven to call 120.

7. A rear passenger safety monitoring system based on facial recognition as claimed in any one of claims 1 to 4, characterized in that: The monitoring system also includes a rear passenger position detection module, which is used to detect whether the rear passengers are within a set safety range. Its output end is connected to the cockpit domain controller. After the cockpit domain controller detects that the rear passengers are beyond the safety range and the duration reaches a set time threshold, the cockpit domain controller displays the delayed alarm time through the on-board large screen and dials the alarm number after the delay ends. At this time, a cancel alarm button is provided through the on-board touch screen or hardware button. If no cancellation signal is received after the delay is completed, the alarm number is directly dialed or an alarm text message is sent.

8. A rear passenger safety monitoring system based on facial recognition as claimed in any one of claims 1 to 4, characterized in that: The cockpit domain controller is provided with a dedicated memory for storing face recognition data of rear passengers. The face recognition data is encrypted and saved and deleted after a set time.

9. A rear passenger safety monitoring method based on facial recognition, characterized in that: The facial image data of the passengers in the rear seats is collected through the camera; the cockpit domain controller identifies the status of the rear passengers based on the facial image data and sends a reminder message to the driver through the reminder module based on the status of the rear passengers; when it is identified that the rear passengers are in an abnormal state, the cockpit domain controller drives the large-screen display of the vehicle to display the rear images taken by the camera and reminds the driver through voice that the rear passengers are in an abnormal state.

10. The rear passenger safety monitoring method based on facial recognition as claimed in claim 9, characterized in that: After the driver learns of the abnormal state of the rear passenger through the in-vehicle large screen and voice, the cockpit domain controller monitors in real time whether the call button is triggered. When the call button is triggered, it drives the in-vehicle phone module to call 120.

Images